Razi University

  With the rapid expansion of fifth-generation mobile communication networks (5G), the increasing number of users and growing traffic volume have made energy consumption management at base stations one of the major challenges in modern telecommunication networks. This study investigates the problem of energy consumption optimization in 5G networks, with a particular focus on base station management while maintaining Quality of Service (QoS). To this end, the performance of five approaches, namely Baseline, Power Control, Sleep Mode, Hybrid, and RL-Based, based on the Q-learning algorithm, is evaluated and compared within a common simulation framework. The simulations are conducted in MATLAB, and the performance of the methods is evaluated based on power consumption, energy efficiency, SINR, and the average data rate of users. Furthermore, to investigate the performance of the methods under different conditions, a Monte Carlo simulation with 100 independent runs is performed, and the resulting data are statistically analyzed using   . The results indicate that the Hybrid method provides favorable performance in terms of reducing power consumption and improving energy efficiency, while the Baseline method achieves the highest average data rate. The RL-Based method also demonstrates competitive performance compared with the other approaches by reducing energy consumption while maintaining an appropriate data rate. Overall, the findings indicate that effective energy management in 5G networks requires a balance between reducing power consumption and maintaining an acceptable level of Quality of Service (QoS). Keywords: Energy Consumption Optimization, 5G Networks, Base Station, Power Control, Sleep Mode, Hybrid Method, Reinforcement Learning, Q-Learning.